justanotheruser has quit [Ping timeout: 268 seconds]
srk has quit [Ping timeout: 240 seconds]
AlwaysLivid has quit [Remote host closed the connection]
srk has joined #nixos-dev
AlwaysLivid has joined #nixos-dev
mkaito has quit [Quit: WeeChat 3.0]
orivej has joined #nixos-dev
<elvishjerricco>
Working on my from-scratch initrd again: Ok I've got the output of makeModulesClosure in /lib/modules, and I've got udev starting with the default rules shipped with systemd in /lib/udev. But the drive still doesn't show up in /dev
orivej has quit [Ping timeout: 246 seconds]
srk has quit [Ping timeout: 240 seconds]
srk has joined #nixos-dev
jonringer has quit [Ping timeout: 260 seconds]
AlwaysLivid has quit [Remote host closed the connection]
AlwaysLivid has joined #nixos-dev
Jackneill has quit [Ping timeout: 240 seconds]
kalbasit_ has quit [Ping timeout: 240 seconds]
Jackneill has joined #nixos-dev
peelz has quit [Ping timeout: 240 seconds]
AlwaysLivid has quit [Ping timeout: 240 seconds]
AlwaysLivid has joined #nixos-dev
qyliss has quit [Quit: bye]
qyliss has joined #nixos-dev
cole-h has quit [Ping timeout: 265 seconds]
teto has joined #nixos-dev
<pie_>
I just sent myself on a wild goose chase because nix-shell shell.nix -A nixpkgs.someattrs doesnt resolve the attr from shell.nix but from the global nixpkgs
AlwaysLivid has quit [Remote host closed the connection]
dstzd has joined #nixos-dev
dstzd has quit [Remote host closed the connection]
teto has joined #nixos-dev
dstzd has joined #nixos-dev
mkaito has joined #nixos-dev
mkaito has joined #nixos-dev
mkaito has quit [Changing host]
dstzd has quit [Client Quit]
dstzd has joined #nixos-dev
dstzd has quit [Client Quit]
dstzd has joined #nixos-dev
WilliButz has quit [Remote host closed the connection]
WilliButz has joined #nixos-dev
orivej has quit [Ping timeout: 264 seconds]
<infinisil>
Problem: fetchurl/fetchzip/etc. reuse an existing path if it has the same hash, even if the url changed
<siraben>
infinisil: is there a way to prevent that?
<infinisil>
Suggestion: Embed the inputs of these calls into the derivation name, so that when they change, the output path changes, so paths with different inputs can't be reused
<infinisil>
Variant 1: Embed the url itself into the name, e.g. /nix/store/smr6clfp2sy15qxxs15r5mqv31cj771k-https-example.com-some-url.tar.gz
tilpner has quit [Remote host closed the connection]
<infinisil>
Advantage: It shows you what it is. Disadvantage: It could be really long
<infinisil>
Variant 2: Embed a (potentially truncated) base64 hash of the url into the name, e.g. /nix/store/srns89ci0mpafjgja3jh20is8ij083c1-8pmJ2TYkgXpiYdwGUc0=
<infinisil>
Advantage: It's the same length regardless of url. Disadvantage: It doesn't show you what it is, and it could be a security problem if the hash is truncated
<andi->
that breaks all kinds of mirror usage
<infinisil>
Variant 3: A mix of 1 & 2: Embed a truncated hash into the name, plus the basename of the url (which is what it does currently)
<infinisil>
andi-: How so?
<infinisil>
(and which variant)
<andi->
oh well I guess it doesn't as it happens during runtime
<andi->
the nix expressions just sees mirror://… right?
<infinisil>
Yea
tilpner has joined #nixos-dev
<andi->
Then ignore me
tilpner has quit [Client Quit]
tilpner has joined #nixos-dev
<infinisil>
Maybe the best solution is to not truncate the hash at all, so that you have a full sha256 in base64, like 9xrcsMDds2S/90wOvz1yPkzn1CIjDv06EzfpuKHeAr4, plus append the basename of the url
<infinisil>
Any opinions on this? Or any better suggestions?
<infinisil>
If not, I'll probably draft up an RFC for this
tilpner has joined #nixos-dev
orivej has joined #nixos-dev
orivej has quit [Ping timeout: 260 seconds]
jonringer has joined #nixos-dev
mkaito has quit [Quit: WeeChat 3.0]
orivej has joined #nixos-dev
<ma27[m]>
gchristensen: in case you have sufficient time and motivation, would you mind looking at https://github.com/NixOS/hydra/pull/743? I addressed a few of the comments now and implemented a simple migration script.
<{^_^}>
hydra#743 (by Ma27, 35 weeks ago, open): Add a filter for maintainers in the jobset-eval view
<lukegb>
infinisil: hmm, I feel slightly -0 about this idea, mostly because I like that e.g. if the source URL changes because upstream moves things around then it doesn't force a rebuild
<lukegb>
but this is a minority of cases, admittedly
<infinisil>
lukegb: Also, the hash will stay the same. And with content-addressed-derivations coming soon, this would even abort early with a rebuild because of that
<lukegb>
Sure, the hash will stay the same
<lukegb>
but there's something elegant about not having a rebuild for "the same" content even if it's coming from somewhere else
<infinisil>
Hm yeah, point taken, but yeah that's a very minor use-case :)
<lukegb>
I think this would also break the factorio package (although presumably it could just reimplement a fetcher) because it relies on you either providing a secret or prepopulating the thing into the store with a relatively obvious name
<infinisil>
Overriding the name would still work, which seems to be what factorio does
<lukegb>
Yeah. In any case, my objections aren't strong enough for me to bother voicing them on the RFC issue
<infinisil>
I guess I should mention in the RFC that this would only be the default `name` if it's not overridden
<infinisil>
I think even with minor objections, the RFC (or anything like it) is still far worth it, because it trips up so many users
<infinisil>
Probably every Nix user at least once actually
* lukegb
nods
rajivr has quit [Quit: Connection closed for inactivity]
b42 has quit [Ping timeout: 260 seconds]
saschagrunert has joined #nixos-dev
<supersandro2000>
Is anyone interested in my nixpkgs-review prompt hook? If so I would try to extract it from my dotfiles and put it in its own repo
saschagrunert has quit [Client Quit]
v0|d has joined #nixos-dev
saschagrunert has joined #nixos-dev
saschagrunert has quit [Remote host closed the connection]
saschagrunert has joined #nixos-dev
b42 has joined #nixos-dev
peelz has joined #nixos-dev
cole-h has joined #nixos-dev
<supersandro2000>
currently it runs nixpkgs-hammering on all changed packages, greps for 0 tests run in pytest and currently I am building that it checks also hydra for new and already failing builds
Baughn has quit [Quit: ZNC 1.6.2+deb1 - http://znc.in]
<cole-h>
Don't really have a dog in the race, so to speak, but I agree with you, infinisil.
Jackneill has quit [Ping timeout: 260 seconds]
<infinisil>
Hm yeah, I just don't know how to do it better
<infinisil>
There's not even any way to modularize such a wrapper command
<cole-h>
Maybe it could just be `optionalString (cfg.proxy != null) "${cfg.proxy} "` and the user sets `proxy = "${pkgs.proxychains}/bin/proxychains4";`?
<cole-h>
Of course, that would only support proxies that allow running a specific command under the proxy as the proxychains example does
<elvishjerricco>
Ok, I've got systemd-modules-load.service working, systemd-udevd.service working, and yet there are still no block devices
<infinisil>
cole-h: Hm yeah, considering that proxychains can do it though, it's probably possible to make all other proxys work like that too with some wrapper if needed
Jackneill has joined #nixos-dev
<samueldr>
I have an issue under pretty specific conditions, where a given .drv fails (reproducibly enough?) on nixos.org hydra builders, but not on my own setup